kecal / tokenizer_config.json
danielsebesta's picture
Release Kecal v1.0 (Round 5: KSK + Vystadial + full subtitles)
9fe3d3c verified
Raw History Blame Contribute Delete
817 Bytes
{
"added_tokens_decoder": {
"0": {
"content": "<pad>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"2": {
"content": "<unk>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"3": {
"content": "<turn>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"bos_token": "<turn>",
"clean_up_tokenization_spaces": false,
"eos_token": "<turn>",
"extra_special_tokens": {},
"model_max_length": 128,
"pad_token": "<pad>",
"tokenizer_class": "PreTrainedTokenizerFast",
"unk_token": "<unk>"
}